I have selected for this project the dataset about bad drivers in U.S. I pretend to visualize which State has the worst drivers in the country and if the number of accidents where drivers who were not distracted is considerable. I am not the best driver but I drive without exceeding the limit of speed, and I am always afraid of crashes because Floridian’s drivers are really racing and that scares me a little bit.

Getting the data

library(tidyverse)
## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.0 ──
## ✓ ggplot2 3.3.2     ✓ purrr   0.3.4
## ✓ tibble  3.0.4     ✓ dplyr   1.0.2
## ✓ tidyr   1.1.2     ✓ stringr 1.4.0
## ✓ readr   1.4.0     ✓ forcats 0.5.0
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()
library(htmlwidgets)
library(ggplot2)
bad_drivers<-read_csv ("https://raw.githubusercontent.com/reisanar/datasets/master/bad_drivers.csv")
## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   state = col_character(),
##   num_drivers = col_double(),
##   perc_speeding = col_double(),
##   perc_alcohol = col_double(),
##   perc_not_distracted = col_double(),
##   perc_no_previous = col_double(),
##   insurance_premiums = col_double(),
##   losses = col_double()
## )

This dataset contains the following variables:

Summarizing the data

After getting the data, it is time to formatting the data because the columns “perc_speeding”, “perc_alcohol”, “perc_not_distracted” and so on, are percentages of “num_drivers (number of drivers involved in fatal accidents)”. I will mutate some new columns in the following step:

bad_drivers<-bad_drivers%>%
  mutate(speeding_drivers=(num_drivers*perc_speeding)/100)%>%
  mutate(alchol_drivers=(num_drivers*perc_alcohol)/100)%>%
  mutate(notdistracted_drivers=(num_drivers*perc_not_distracted)/100)%>%
  mutate(noprevious_drivers=(num_drivers*perc_no_previous)/100)

Plotting the data

Below will be a Treemap that will show us the number of bad drivers in U.S grouping by State.

library(ggplot2)
library(treemapify)
treemap1<- ggplot(bad_drivers, aes(area = num_drivers, fill = state, label = paste(state, num_drivers, separate = " "))) + 
  geom_treemap(show.legend = FALSE) +
  geom_treemap_text(fontface = "italic", colour = "white", place = "topleft", grow = FALSE) +
  labs(title = "Number of Bad Drivers in U.S", subtitle = "Per State")
treemap1

I will create some barplots for analyzing the data and see what are the major percentage of drivers involved in fatal collisions, and which state has the most of them.

library(plotly)
## 
## Attaching package: 'plotly'
## The following object is masked from 'package:ggplot2':
## 
##     last_plot
## The following object is masked from 'package:stats':
## 
##     filter
## The following object is masked from 'package:graphics':
## 
##     layout
speedbarplot<-bad_drivers %>% 
  select(state, num_drivers, speeding_drivers) %>% 
  gather(type, value, num_drivers:speeding_drivers) %>% 
  ggplot(., aes(x = state,y = value, fill = type)) + labs(fill="Type") +
  geom_bar(position = "stack", stat="identity") + 
  scale_fill_manual(values = c("green", "darkgreen")) + xlab("States") +
  ylab("") + theme_minimal() + theme_classic() + 
  theme(axis.text.x=element_text(angle=90, vjust=0.5)) 
ggplotly()
speedbarplot

As we can see here, for speeding drivers involved in fatal collisions the State of South Dakota, Pennsylvania, West Virginia, Montana and Hawaii have the most speeding drivers involved in fatal accidents in the U.S.

alcoholbarplot<-bad_drivers %>% 
  select(state, num_drivers, alchol_drivers) %>% 
  gather(type, value, num_drivers:alchol_drivers) %>% 
  ggplot(., aes(x = state,y = value, fill = type)) + labs(fill="Type") +
  geom_bar(position = "stack", stat="identity") + 
  scale_fill_manual(values = c("green", "darkgreen")) + xlab("States") +
  ylab("") + theme_minimal() + theme_classic() + 
  theme(axis.text.x=element_text(angle=90, vjust=0.5))
ggplotly()
alcoholbarplot

It is annoying how many people drive when they have been drinking, this plot shows that we need to create more conscience about the risk of driving under the influence of alcohol. North Dakota, South Carolina, Montana, West Virginia and Arkansas has the most number of bad drivers involved in fatal collisions in this condition.

notdistractedbarplot<-bad_drivers %>% 
  select(state, num_drivers, notdistracted_drivers) %>% 
  gather(type, value, num_drivers:notdistracted_drivers) %>% 
  ggplot(., aes(x = state,y = value, fill = type)) + labs(fill="Type") +
  geom_bar(position = "stack", stat="identity") + 
  scale_fill_manual(values = c("green", "darkgreen")) + xlab("States") +
  ylab("") + theme_minimal() + theme_classic() + 
  theme(axis.text.x=element_text(angle=90, vjust=0.5))
ggplotly()
notdistractedbarplot

Here the State of West Virginia, Montana, South Carolina, North Dakota and Arkanzas has the most number of fatal collision were not distracted drivers were involved. So, the chance of being involved in a fatal collision when you are not distracted are minimum comparing that with being speeding or alcohol-impared. Now, let’s see what happens with people that haven’t been involved in fatal collisions before.

nopreviousbarplot<-bad_drivers %>% 
  select(state, num_drivers, noprevious_drivers) %>% 
  gather(type, value, num_drivers:noprevious_drivers) %>% 
  ggplot(., aes(x = state,y = value, fill = type)) + labs(fill="Type") +
  geom_bar(position = "stack", stat="identity") + 
  scale_fill_manual(values = c("green", "darkgreen")) + xlab("States") +
  ylab("") + theme_minimal() + theme_classic() + 
  theme(axis.text.x=element_text(angle=90, vjust=0.5)) 
ggplotly()
nopreviousbarplot

Let’s see what happens with alcohol drivers that are involved in fatal collisions and insurance premiums.

plot1<-ggplot(data=bad_drivers, mapping = aes(x=alchol_drivers,y=insurance_premiums)) + 
  geom_point() + geom_smooth(method="lm")+ 
    labs(title = "Insurance Premiun vs Alcohol Drivers",
       x = "Drivers under influence of Alcohol",
       y = "Insurance Premiums")+
    scale_y_continuous(labels = scales::comma)+
  theme(legend.title = element_blank(), legend.position = "none") + theme_minimal()
ggplotly()
## `geom_smooth()` using formula 'y ~ x'
plot1
## `geom_smooth()` using formula 'y ~ x'

The relationship between alcohol drivers and premium insurance are indirectly proportional. Insurance companies tends to claim your DUI was intentional, and they can refuse to cover you or drivers will face higher insurance premium.

Map

Below is the map that shows the U.S states with bad drivers percentage.

library(plotly)
library(ggrepel)
library(sf)
## Linking to GEOS 3.5.1, GDAL 2.2.2, PROJ 4.9.2
library(usmap)
baddrivers_map<- plot_usmap(data = bad_drivers, values = "num_drivers", color="red") + 
  scale_color_continuous(name = "Bad drivers in U.S", label = scales::comma) + 
  theme(legend.position = "right")
baddrivers_map